Writing, Editing, and AI with Heidi Nobles

Back in August, I had the opportunity to hear a short presentation from Heidi Nobles, assistant professor in writing and rhetoric and director of Writing Across the Curriculum at the University of Virginia. The presentation was part of a two-day institute on teaching and generative AI, and Heidi leveraged her background as an editor to provide a different way of thinking about working with generative AI.

Heidi pointed out that when we ask ChatGPT or some other AI chatbot to polish a draft essay, we’re asking for copyediting. That’s useful, yes, but there are other, earlier stages to an editing process. Might AI be useful during those other stages? Heidi argued for yes. A chatbot won’t be as good as a human editor, but most writers don’t have access to a human editor, so it’s worth exploring what AI can do.

On today's podcast, Heidi Nobles talks about writing and teaching writing from an editor's perspective.

Episode Resources

·       Heidi Nobles faculty page, https://wac.virginia.edu/people/heidi-nobles 

·       Edits on the Record, https://editsontherecord.com/ 

·       Choose Your Own Adventure maps, https://www.cyoa.com/pages/choose-your-own-adventure-these-maps-reveal-the-hidden-structures-behind-the-books 

·       One Book, Many Readings by Christian Swinehart, https://samizdat.co/cyoa/

Podcast Links:

Intentional Teaching is sponsored by UPCEA, the online and professional education association.

Subscribe to the Intentional Teaching newsletter: https://derekbruff.ck.page/subscribe

Support Intentional Teaching on Patreon: https://www.patreon.com/intentionalteaching

Find me on LinkedIn and Bluesky.

See my website for my "Agile Learning" blog and information about having me speak at your campus or conference.

Transcript

Derek Bruff (00:05):
Welcome to Intentional Teaching, a podcast aimed at educators to help them develop foundational teaching skills and explore new ideas in teaching. I'm your host, Derek Bruff. I hope this podcast helps you be more intentional in how you teach and in how you develop as a teacher over time. Back in August of 2024, I had the opportunity to hear a short presentation from Heidi Nobles, assistant professor in writing and rhetoric and director of writing across the curriculum at the University of Virginia. The presentation was part of a two day institute on teaching and generative AI and Heidi leveraged her background as an editor to provide a different way of thinking about working with generative AI. When I hear editing as in editing an academic paper or a book, I usually think of copy editing, but of course there are other forms of editing. Heidi pointed us to substantive or subed editing, which involves working with the structure and organization of a manuscript and developmental or dev editing, which involves helping an author shape their ideas even before there is a manuscript.

(01:09):
Heidi pointed out that when we do the common task of asking chat GPT or some other AI chat bot to polish a draft essay, we're asking for copy editing. That can be useful, but might generative AI be useful during the substantive or even developmental editing stages of a writing process? Heidi argued for, yes, a chatbot won't be as good as a human editor, but most writers don't have access to a human editor, so it's worth exploring what AI can do. I invited Heidi on the podcast to share her perspectives on editing and AI, and I'm excited to share that interview today. Two things to know before we get to the interview. One is that Heidi mentions a project she's leading called Edits on the Record. In this project, she and her team are making public the entire editing process of a new book as a way to make that process more visible. The second is that we get to the topic of working with students on using AI in their writing, but not until the end of the conversation. we'll camp out on our own academic writing and the editing roles that AI might play in it for eventually bringing those ideas into the classroom. But don't worry, we will get there.

(02:15):
Thank you, Heidi, for coming on the podcast today. I'm very excited to talk with you and to get to know you a little bit better. Thanks for being here.

Heidi Nobles (02:23):
Thank you so much for having me. I'm excited to join your program.

Derek Bruff (02:27):
And I'll start with my usual opening question. Can you tell us about a time when you realized you wanted to be an educator?

Heidi Nobles (02:35):
So I went to a very small undergrad, very small private school in the middle of Oklahoma. Not very elite institution, but I was lucky to be in their experimental honors program. I was in the third class and there were only 40 of us. And those faculty, all of our classes were cross-listed. They were all co-taught and interdisciplinary, and the faculty were really invested. So they had us over to their houses for dinner. They would dress up and serve pancakes at midnight during finals week to all the students, and I still go back and talk to them today. They were just really, really good mentors, and that four years was a celebration of knowledge. It was all the things that education's supposed to be. I mean, I was a nerd clearly, but I was able to be a happy nerd there with all of my peers, and I didn't realize how much that had shaped me, I think, until I was in my master's program.

(03:37):
So I had gone at the recommendation of one of my mentors to get my master's in literature, didn't know what that really meant or what I was going to do with that. And again, this will come up in a little while. At the end of my second year, I interned at our university press. Really liked that kind of felt professionally that was a good fit for me, and yet then the next semester I had to teach as part of my assistantship was not really looking forward to that. And about two or three weeks in realized, oh, no, I really liked this. I really liked teaching writing, and I like working with the students and all of my, I mean, I could just hear all of those mentors in the background coming back to me and that this was my time to pass that on. And so yeah, I fell in love with both of them at the same time. I wasn't planning to teach, but I've really enjoyed it and for me, teaching, writing and working and editing has always been intertwined.

Derek Bruff (04:33):
Wow, okay. I'd love to explore that connection a little bit, but let's talk about editing. We're going to get to artificial intelligence here in a minute, but I do think you have a lot of experience as an editor and you've made some interesting connections between using AI and editing. When many of us hear about editing, perhaps especially our students, but even us as academic professionals, we think of copy editing. I gather that's only a small part of the editing process. Can you walk us through the different types of editing that are in the field?

Heidi Nobles (05:15):
Yes. So this is this foundational conversation for me. I feel like every time I give a presentation, I have to have a preparatory. When I talk about editing, what I mean is, and yes, so copy editing, very important, but one idea I'll probably come back to is, and it's not very, I mean it's profound, but it's not original to me. The idea that writing is thinking. So when it's a form of thinking and it lets us do it, and so when we limit that to copy editing, we're limiting our ability to think in a wider range. So I'm in the English department, we spend a lot of time worrying about the sentence level and why it is incredibly important for refining those clauses and making sure that the vocabulary is correct and all of the precision is important for thinking. And also for me, editing starts way before that. Developmental editing tends to be more, it might even be before the paper, the paper or the book, whatever it is, even written down at all. I might be, someone might call me and say, I have an idea and I dunno how to turn it into a book. And we go out and we have coffee and we talk about it, and that is a stage of the developmental edit.

(06:24):
The most recent job that I had as a freelancer was as a developmental editor for Manning Publications, which is a tech publisher. And that one was fascinating because there aren't very many houses who have dev editor as an in-house job where it's not just a totally outsourced thing. They had an editorial manager who had a whole philosophy of developmental editing. And so there was a team of us as dev editors who worked with these authors. And I don't know, it's a weird construct of a publisher anyway, because most publishers that I've worked for, the authors are sending in all of these manuscripts and proposals and they really want us to take their books. And in this company, these people don't want to be authors. The acquisitions editors, it's their job to, they don't work with the text. They just go out and get the books, and it's their job to find these tech people who are doing interesting things and then talk them into creating a book for us. And so what they do is pair them with a dev editor like me to say, Hey, this person has no idea how to write a book, but they did this cool thing, walk them through it, and then it's my job to drag the book out of them. Okay.

Derek Bruff (07:36):
Yeah.

Heidi Nobles (07:38):
And so anytime you're working with an author on ideas in progress, and maybe the manuscript is partially written, but it's not like you have an actual draft in front of you that you're changing the sentences on most of the time. So that would be developmental editing, and then sub editing is organization and structure for the most part. So now you probably have a full manuscript, or at least mostly complete and you're moving things around. So you might be saying, Hey, you've got this whole section devoted in chapter six to a topic that no one cares about, but in the meantime, everyone's really curious about what you said in chapter two, and then you just left it hanging, please fix that. Or We don't understand what you're talking about here because you don't explain it until chapter 12. Let's move those around. But big picture stuff that you're moving around, and then once you finish all of that, then you can deal with copy edit, but there's no point in changing the sentences before the ideas are developed or before the whole structure of the thing Makes sense.

Derek Bruff (08:36):
I'm just reflecting on my own experience as an author, and I was in the category where I felt like I kind of knew the book that I wanted to write, so I came in with fairly firm ideas, but then even then did some developmental editing, at least for my second book,

(08:54):
I worked with the editor to really think about who is the audience here and what do they really need? And of these things, do you want to talk about Derek? Which are the ones that would really connect with that audience and which are your pet topics that no one cares about? It's that kind of work. And then I think for the substantive editing, I got a lot of that from the peer reviewers on my

(09:16):
Book Where they were able to point out, Derek, you're going on and on in the introduction about what the book is not You're trying to caveat your life. Just tell us what it is and then move on. Don't worry about what it's not. Or chapter one is really thin, what happened there? That was really helpful feedback for me. Well before the copy editing stage.

Heidi Nobles (09:37):
Yes.

Derek Bruff (09:40):
Okay, so I want to take this to the classroom because this thing, we've both observed that people hear editing and they think copy editing. I think that's very true of the writing students that I've taught, right?

Heidi Nobles (09:55):
Yes.

Derek Bruff (09:55):
I asked them to revise something and they're making very much sentence level changes. So what are some strategies you've used or you've seen to help students move from that copy editing to really kind of embrace the developmental or substantive editing?

Heidi Nobles (10:14):
So straightforward would be don't give them copy edits, which my first few years of teaching, especially I was in a literature program, we were focused so much on the fine granular or items at the level of the sentence, and I cared so much, and I do, I care about it. I love a good copy of it, but I would go through, I was trying to perfect their papers, and one thing we say in composition studies is we're not here to perfect their papers. We're here to equip them as writers, and if I perfect their paper, it's a kind of disposable document. It's supposed to train them to do something, but if I fix it for them, they're not learning it. So our best practices in composition studies is don't give them very many line edits because that is what they'll focus on. The other issues will feel overwhelming and they'll just zero in on the lowest hanging fruit. Also, if we give them too many comments, it's overwhelming and they can't process it all. And we're working at multiple levels, and I guess this is where my language will overlap because this isn't how we talk about it in comp studies, but it's how I talk about it as an editor, and it also gets into ai it,

(11:28):
But a document goes through all these different stages and also functions at multiple levels. And so if we are giving students developmental feedback and style feedback at the same time, they can't do that. That's not how it works. And so what I do is if there's a copy edit that genuinely matters for sense. I had a student recently who had said something about a derivation in her first draft, and that made sense. She was talking about a math person, and so I didn't catch it. The first two rounds through, I do a lot of drafts with my students, and then finally in the third one I was like, oh, you don't mean derivation, you mean deviation. And that changed the whole trajectory of the paragraph because that really was a concept change. So then I'll address that, but if it is their periods are outside of their quotation marks or something, no, we can fix that later.

(12:26):
And so I, I have my grading systems set up in stages, and if it is a D paper, if it's in the developmental category, there's no shame in that, but I'm not going to give you any copy edits, no point. So I'm going to give you conceptual comments, and then I give them a few comments about the ideas and they come in and talk to me. If it is at a structural level, which is usually a C paper for me, then that's the only kind of comments I give them. I only talk to them about reorganizing the paper, and then if it's a B paper, that's usually because the grammar's off and then I give them copy edits and that's great.

Derek Bruff (13:04):
So part of what I'm hearing is sequencing, right?

Heidi Nobles (13:07):
Yes.

Derek Bruff (13:08):
So don't lead with the copy edits, but more specifically, figure out where the student is and what they're going to need and don't go past that stage of the process until if it's the ideas, then you need to talk about the ideas. If it's the structure, then you need to talk about the structure, but don't try to accelerate that

Heidi Nobles (13:29):
And they will do the latest stage possible. So if you talk to 'em about structure and they have idea issues, they won't go backwards and they're trying to give you what you want, but they're confused. And developmental and organizational edits are hard as we know from writing. And for the most part, they've never been asked to do them before, and so they've never seen it modeled for themselves before.

Derek Bruff (14:00):
Sure.

Heidi Nobles (14:02):
And so they're usually, in my experience, really uncomfortable at first, and then it's their favorite thing about the class. At first, they're just like, what is this? And does this mean I'm stupid? And no, this is how professionals write. This is a good thing. Then once they realize that they can play a little bit more and they don't have to have the nitpicky editor in their head saying, you split your incentive. If they even know what that means, then they're much, they start having fun playing with the ideas. So when I first started teaching, I did grammar level edits, and when I go back, those were all hard copies. So I can still go back and occasionally see some of the papers and the revisions that I got were terrible. They never changed their mind about anything. I fell into the trap that I think a lot of us do, which is you can't explain to them why their paper is grammatically fine, but it's a D, this is a bad paper. How do you explain that to these poor students? But my sentences are fine. I don't have any. Once I moved to this model, they got it. And to say, yes, this is superficially fine, but there, let's talk about that. They get really excited to have their ideas taken seriously, and now I get dramatic rewrites. It's kind of beautiful.

Derek Bruff (15:27):
That's great. That's great.

Heidi Nobles (15:28):
It's really fun to see what they Do.

Derek Bruff (15:29):
Yeah. Well, I can imagine for some students it's very freeing to say, I'll worry about the grammar later. Right now, I actually get to play with ideas. And I think sometimes it's because particularly in writing across the curriculum that those of us who aren't trained to teach writing and have disciplinary content that we think is really important, I think there's a temptation to have students write about things before they've had a chance to really think about things and come up with their own ideas and hypotheses and conjectures, and we say, Hey, we're going to write a paper on this unit. And they're like, okay, I have to come up with a thesis. And I'm like, when I write, the thesis sometimes comes out halfway through the process when I finally figure out what I want to say. Right,

Heidi Nobles (16:16):
Exactly.

Derek Bruff (16:17):
And so moving them away from the copy editing to these other stages, I think I can imagine that being really powerful. I do want to talk about generative AI because I think you've got a way of thinking about how we can work with generative AI that draws on this rich experience as an editor, and I'd love to hear you kind of talk us through that.

Heidi Nobles (16:37):
So I feel like I have to have the disclaimer of I understand why some people find it terrifying, and I occasionally also have existential crises.

(16:49):
However, yeah, I find it really helpful, and I do think I find that I approach it in my experience differently than most of my peers are because of the editorial background. So again, if we focus on the copy editor, if that's the only thing we let ourselves do, we're limiting our full range of how we use the AI as well. And so very practically, if you're thinking about editing just as copy editing, you're likely to pull up whatever your preferred gen AI platform is, dump your text in and then say, can you clean up my grammar? Or eventually, I don't think they're very good at style guide yet, but eventually it would be great to be able to say, can you make sure this conforms to Chicago 15 or whatever we want to do? That'd be awesome. I can't get it to do that very well yet, but faculty might benefit from experimenting anymore by using AI as a dev editor or sub editor.

(17:43):
So again, practically a few ideas that I use would setting up a beta test group. So you can open your AI agent and describe the people that you would want in your beta test group and ask and then give it text and say, Hey, can you discuss this? And then you can insert yourself in there and say, Hey, I'm your focus group leader, and I was just wondering if you could talk more about this idea or that idea. And those could be just ideas that you're still playing with. It could be, so you could just describe it, Hey, I'm thinking about working out this idea and it can be really, really raw, which is really nice. And then you can see where they go with it and bounce things back and forth and try to prompt them to get the material that you get the responses that you want, which is a simulation of what we would do in a workshop of trying to submit our work.

(18:33):
And in a creative writing workshop, traditionally I submit my work a few days before the class period, and then everybody reads it, and then we go in and I do not get to talk. I just sit there and take notes, and then they discuss my work and I get to see how they responded. And so yes, a passage that I was trying to make them cry and they think it's hilarious, that's a problem, and I don't argue with them. I'm just like, oh, I need to revise that. So similarly, you can try to get that feedback and then you can go back and revise it and see if you get a different response from the virtual beta testers also. That's one idea. Another one is to set up an agent as a peer editor. So if you're sending out an article for review, you can put in the journals review guidelines of what you're going to be assessed on their rubric, and then you can say, Hey, pretend you are. And then you can ask it to be reader one, reader two, or reader three. You just describe that. So you can say it's probably not very helpful to have the nice person respond to it.

Derek Bruff (19:34):
We all know what reader two is like.

Heidi Nobles (19:37):
Exactly. So you could say, Hey, can you be kind? But critical? You can say, can you just be really harsh? What's the worst thing you can say about this? And then it feels less personal because it's just a bot, but it also can screen out. I took one of my articles that had been rejected, so I already had the feedback and then I ran it through, and the feedback that I got from the bot was very similar from the feedback that I got from the reviewers. So again, it's not perfect, but it can be helpful before it goes out to real people.

(20:09):
And then you can also feed, there's different platforms you can use. One that I'm excited about right now is, so Google Notebooks has a podcast feature you can upload. So you can upload. So what I did was take a project that I've got, which is edits on the record.com, and I put in a bunch of the URLs for individual pages. I can't scrape the whole site, but I put in a few individual pages and I uploaded some PDFs and then ask it to generate a podcast. And it was fascinating. So it makes up these two voices. It's got these two cloned people who talk about your project, and they're very enthusiastic, encouraging. It's the best thing they've heard in a long time, but then the way they represent it can be really telling. So I felt like I had given an interview to a newspaper and they got part of it and then misrepresented my work and that frustration of like, no, that's not what I said.

(21:06):
But then it's so helpful to say, oh, okay, why does it think that? What did I not say? And so there was a whole section that it didn't ever talk about the ideas underlying the book, and it was like, that's the whole point. I didn't say that clearly enough. I need to add a whole page about that on the website. And so the podcast, then it takes 10 or 12 minutes to listen to it, and then you can make your own notes about things that you need to change back in the, I would consider that a developmental edit. And then the fourth thing I would say would be feeding messy draft material into again, the agent of your choice and then asking it to, so I have one larger project that's just a hot mess. Actually no, there's two, I'll talk about one that I'm editing right now.

(21:49):
So I have this one book that I'm editing and it is a mess, and it's got a lot of beautiful moments in it, but it does not have cohesion or movement. And so I broke it apart into multiple sections because it was too big for any one agent. But then I asked it to suggest multiple outlines. So I didn't want it to just give me one rigid one, but so what are different ways we could think about organizing this? And then that helped me think about it for myself. And then I went back and made my third human version where I reorganized it, and that was really helpful. So all of these strategies mirror work that you would actually do with a human editor if you had access to that person. And I also think, I don't know, the luxury of working with an actual trained editor is lovely. If you get it, please do it. But given that most of us are doing it with people who don't have training, working with the block can help us learn strategies that are portable for when we're working with our peers.

Derek Bruff (22:46):
Wow. Okay. So I want to circle back to one of those examples because it's reminded me of an analogy you made when I saw you present back in August,

(22:57):
The Choose Your Own Adventure analogy. I am a child of the eighties. I still have my collection of Choose Your Own Adventure books.

Heidi Nobles (23:05):
I just got a couple of board books for the baby. I'm very Excited.

Derek Bruff (23:07):
Oh, wow. That's great. That's great. So how is editing a Choose Your Own Adventure book and how can that help us think about what we can do with Gene ai?

Heidi Nobles (23:17):
Yeah, so I think it helps when there's a visual. So anybody listening, if you go to the Choose Your Own Adventure website, they have some example maps that you can look at where they have mapped out all the different possibilities of their various books. So I don't like Space Adventure, and if you turn to page 37 and whatever and the branching that happens. And so that I think is pretty easy for us to get our heads around because as you're reading the book, you're aware that there's at least 12 different endings or 37, they probably said on the cover how many different endings there were. And so that you're taking different paths. And that's the fun part. But that's true of any book. It's just that you don't see all of the paths. And so when I, as the editor get a book, especially early in the process, I see all of these maybe versions of the book that are branching out, and then it's my job to help the author choose one specific path and then edit that out.

(24:14):
And so then you as the reader never see those. And then hopefully what you get is a much more readable version so that it's not the hot mess that I'm working with, but it's something that's a pleasure to read and that you get the meaning out of. But with the ai, one thing that I have concerns about is that when you ask it to do some of these things for you now, no, you don't see the possibilities either if you ask it, especially if you ask it to just outline your books for you and not think about it. If you ask it for multiples, that helps. And then you go back and you think about it, there's ways to mitigate the problem. But if you just ask it, Hey, can you generate an outline for me about this topic? And it looks pretty good, but you've missed all of those maybe paths and you've cut off all this possibility that is in you that you're supposed to generate and now you're not going to, and there's no other human who remembers any of 'em.

(25:11):
So one of the things that I am responsible for as the editor is to document all of those possibilities. And we don't purge that for usually 18 months after publication. And so all of those things that we can go back to that book, that's a mess. Right now, I have cut so much, the author's such a good sport. I have cut so much material out of that book, but it's all saved in files. And so when we say, oh, now you just added this new thing that reminds me of this thing you wrote last year, and I can go back and pull it and we can put it back, but if you ask the ai, it doesn't remember any of that, or it's just gone into a training model, so it's not something that you can use.

Derek Bruff (25:51):
But it sounds like it does offer some options for imagining what those hypothetical paths are for whatever piece of writing that I'm working on.

Heidi Nobles (26:00):
Yeah, I think so. I think it can be really generative. I just think you want to think about, I don't know, being thoughtful and being in control that you're the author. So if my students are in a peer group in class and they're working out their project and they get an idea from one of their peers, it's still their project that doesn't take that. The other student is not the author. It's just because they threw out an idea. And so I dunno, keeping that in mind that we're in control and it can be helpful.

Derek Bruff (26:34):
I heard someone make this argument a year and a half ago that I thought was very helpful is that if I think about the book that I wrote, I am the author of the book, but I had significant input from lots of other human beings in various different ways that led to kind of the finished product, the editors, the peer reviewers, the people I interviewed, my thought partners when I was working on drafts of it. All of that is part of the mix. And so thinking of generative AI as one of those contributors that is kind of better at some of those contributions and worse at others. Right,

Heidi Nobles (27:09):
Exactly.

Derek Bruff (27:12):
I think that's a useful lens to think about generative AI in this process and to remind ourselves that in fact, it is a process to begin with that has lots of contributions.

(27:23):
Hi, this is Future Derek breaking in to say that someone who made that argument was Julian Posada, assistant professor of American Studies at Yale University. I heard him give a talk about AI and authorship when I was visiting Yale in 2023.

Heidi Nobles (27:38):
And we're comfortable with that with other kinds of creations. So when we go to a movie and we see the credits roll, it is not shocking to us that there were thousands of people who contributed to this movie, and yet we have this long established myth of the quote author capital.

Derek Bruff (27:55):
Well, it reminds me of something else I heard you talk about back in August, which is as an editor thinking about working with this author, looking at all the possible paths this book could take, the heuristics you're using are not necessarily the heuristics that a chat GPT is going to use. And one of those, I remember, again, mainly because my wife is in publishing, and it made a little sense to me, is that part of what you're thinking about as an editor is the marketing is the physical production limits things about page counts and trim sizes that authors don't know, and that sometimes those things really matter a lot for the shape of a book.

Heidi Nobles (28:35):
Yes, very much. Yeah. And you can probably train a bot to take those things into account, but we don't have an agent that does that.

Derek Bruff (28:46):
Right?

Heidi Nobles (28:46):
Yes. And like you said, even then it would do really well at certain parts of that, and then it wouldn't do well Other,

Derek Bruff (28:53):
What does a chat bot tend to make decisions on, if we can call it making decisions?

Heidi Nobles (28:59):
Yeah, fair enough. Oh, that's a good question. I'm going to think about, so I used it this last week to help me edit a book. So the things that I tried, I divided the manuscript in half so that I could put it in, and I asked it for a few different things. I tried to get it to identify recurring themes and then compared what it thought the themes were. One of my assistant editors, bless her over the summer, went through the whole manuscript many drafts ago, but she color coded it, and she came up with seven or eight themes that were coming through and then matched them for shape and tried to see yes, what on earth was happening. And so I asked the bot to do that with the newer draft and then compared its notes to her notes to my notes, I asked it to break up the chapters because it had six chapters that were all really long and circuitous.

(29:58):
And so now it has 14. The first round was really good. It gave me the first three or four chapter breaks, and it told me, try ending it on this line. And then I went back to that line and I either really liked it or I liked something nearby it to be a chapter break. And again, we've been through enough drafts at this point that it's been pretty good shape for that. But after about four rounds of that, it started getting confused. And so then it was giving me chapter breaks that were either before the last one or just two more sentences later, so I couldn't keep it up. But what I was able to then see from those first four or five chapters was the pacing and how it felt. And it did feel really good. The author's a poet, it's very poetic, which is part of why it's a jumble, because she does much better in beautiful short form. Again, we'd reorganized it so many times. And I think between those, I don't know, somewhere along the way, the chapters now are consistent. They start in sensible places, they end in sensible places, and they have enough of a theme that the chapter titles now make sense, and you can see the trajectory and the movement, and she starts in a place and ends in a place, and you can see the characters growth through the book. Yeah,

Derek Bruff (31:15):
Yeah. Well, let's talk about working with students. How do you work with students to work with AI to make some of these editing moves in a useful way?

Heidi Nobles (31:24):
I will say that for this semester, what I tried was a three round arc. So I gave them a foundations of AI literacy and what I want them, which was a lot of actually what I talked about at that August seminar. I was just like, we're afraid of learning loss. We don't want you to just use these shortcuts. What are trying to accomplish? Think about your own goals as a learner and as a writer, and then use the ai, ask yourself those questions, and that will help guide your own decision making in terms of heuristics of how to use the different ais that are now available and will become available because we can't anticipate at all. So we did that class and I just injected these into a bigger class period. So we did about 20 minutes on that, and then we went on, and then I did 10 minutes into future class periods, and one was using AI as a dev and subedit, and here are some specific things that I want you to try.

(32:17):
And then the next one was, how do you use it for style? I showed them some of my own, I think I pulled up a conference proposal that I submitted this semester, and I had written the whole thing. I spent a couple hours on it, and it was stuff I knew, but then I could tell that the sentences were really long. I knew they were framed in ways that I found more precise, but that readers would find very frustrating. And I couldn't get myself to change it just morally, but I knew the bot would have no moral qualms with changing it. And so I fed it in and said, can you revise this for readability, assuming a scholarly audience and not changing any of the ideas. And it gave me back another version. I put both of those up on the screen next to each other, and then I kind of walked them through how I chose which edits to accept and which to reject and why.

(33:15):
And then I asked them to try that on their own. And that seems to be helpful. I think the students really benefit from very practical experimentation and guidance like mentoring in the moment. So mine have been much more transparent with me since then. I think they're using it more than they were before, but in a good way, they are citing it. So they've got lists of the prompts that they used. And that's been, I think actually that's going to be really educational for me to do a meta review afterward of the kind of prompts that they are using, because some of them are helpful, some of them are not. And we do the Cuse survey every year. And our institutional research and analytics office was kind enough to send me the three questions from last spring that were about ai, AI perceptions and use among our undergraduates.

(34:09):
And the number one thing that the students were using it for was research. 64% of our students were using it for research, and I do not want them using it for research because it's making stuff up. I would be happy for them to use it for editing. So talking to them about, I would rather you use Wikipedia by far, because those are static. Somebody has looked at it, they're vetted, and the other one is just predictive human text. So talking to them about that and why you would do it one thing and not the other. But that all comes through, I guess we're back to the apprenticeship model, which I hate, but I mean, I love it. I hate the privilege of it. And then who's excluded? So trying to make that transparent for other people too.

Derek Bruff (34:55):
There's been a lot of conversation in higher ed, particularly in writing instruction circles about how writing instruction needs to change in light of chat GPT. Some folks are wanting to double down on what we were doing five years ago. Some folks are wanting to change everything up. John Warner is quite famous for saying What we were doing five years ago was crap anyway. So this is a great opportunity to actually teach writing as thinking. Do you think writing courses should have different goals now that we have access to tools like chat, GPT and copilot?

Heidi Nobles (35:34):
I don't think we have different goals, but I guess that on what the goals were before, my goals have not changed because my goals were to teach students how to use writing as a thinking tool and a communication tool. And that doesn't change, but the heuristics and how we do it changes dramatically. I think, not right away, but over time. And I think with our pedagogy, we have to adapt how we teach the things that we are still trying to convey. And maybe a useful, I don't know if this is useful or not, but analog would be. So in an introductory coding class, the classic first assignment has been for years world, right? Teach your students how to code hello world and whatever the programming language is and get a print statement. And now, when I took that class, I had to really, it's an introductory class, but I had to think about it, what variables am I pulling here and what are the commands and what syntax am I supposed to be using?

(36:34):
And now I can type it into ChatGPT, and it's less than a second and it spits out the answer. And so maybe that assignment as is not helpful anymore, but annotating that code is very helpful. So the students going in and saying, okay, so here is the thing that I'm trying to generate. Here's the code. How did it get from point A to point B? What are all of these things doing? They need to understand that. So I think similarly with us, there are probably ways that we adapt our teaching, but I think there are more creative things that we need to be doing than just trying to get them offline and stop it. That's not going to help. And I think that shuts down the thinking. Not that there's not a place for handwriting, go for it sometimes.

(37:17):
But there are really helpful ways to use the technology. And I think when the faculty members are willing to experiment with it themselves offline, behind closed doors, they will find things that are helpful that they can then talk to their students about. And I also think the more transparent the students are about what they're actually doing in our context, a lot of faculty members will find more confidence that their students are being more creative and thoughtful than they thought. Now, again, there are other places where I'm well aware that the students are just shoving it in there and generating trash, and that's incredibly disheartening. That's a different conversation.

Derek Bruff (37:54):
Yeah. Yeah. Well, thank you, Heidi. This has been really delightful. Thanks for coming on the podcast, and

Heidi Nobles (37:59):
thank You for Having me

Derek Bruff (37:59):
walking us through a little bit of the world of editing. I think it's really fascinating. So thanks for being here.

Heidi Nobles (38:05):
Thank you so much. I'll look forward to talking to you again sometime.

Derek Bruff (38:11):
That was Heidi Nobles, assistant professor in writing and rhetoric, and director of writing across the curriculum at the University of Virginia. Thanks to Heidi for taking the time to come on the podcast and sharing her perspective as an editor on writing and ai. I have to admit, when I saw Heidi present at that workshop back in August, she had me as soon as she put a map of a Choose Your Own adventure book on the screen. I'm a sucker for a good Choose Your Own Adventure reference. Those books brought me so much joy as a child. In the show notes, you'll find links to a couple of different mappings of Choose Your Own Adventure books. You'll also find a link to the edits on the record project that Heidi mentioned during our conversation.

(38:49):
Also, I usually record these podcast episodes to be evergreen that is not tied to a particular time, but hopefully useful for a long time after publishing. But I have to take at least a minute and acknowledge this moment in higher education here in February, 2025, in which so many students, faculty, and staff are quite frankly, scared for their careers or their livelihood or their personal safety thanks to the words and actions of the new US Presidential Administration. I can only hope that when someone listens to this years in the future, that US Higher education and frankly, the United States is in a better place.

(39:28):
Intentional Teaching is sponsored by UPCEA, the Online and Professional Education Association. In the show notes, you'll find a link to the UPCEA website where you can find out about their research, networking opportunities, and professional development offerings. This episode of Intentional Teaching was produced and edited by me. Derek Bruff. See the show notes for links to my website, the Intentional Teaching Newsletter, and my Patreon, where you can help support the show for just a few bucks a month. If you found this or any episode of Intentional Teaching useful, would you consider sharing it with a colleague? That would mean a lot. As always, thanks for listening.